Gradient First-Order Derivative:The gradient, or first-order derivative, measures how quickly the brightness of the pixels in an image vary. It indicates how rapidly pixel values change with respect to changes in position.The gradient is often used for edge detection in image processing. High gradient values are useful for identifying edges since they are typically seen at the borders between objects or areas in an image.Operators like the Roberts cross operator, Prewitt operator, and Sobel operator can be used to calculate an image's gradient.Once the gradient has been calculated, edges in the image can be found and highlighted using edge detection methods like thresholding or edge linking. Laplacian Second-Order Derivative:The gradient's rate of change is determined by the second-order derivative. It gives information on the sharpness or curvature of variations in pixel intensity.An image with strong intensity variation indicates the areas that the Laplacian operator operates on; these regions usually correspond to edges or corners of the original image. Since the Laplacian operator is noise-sensitive, it is frequently combined with smoothing methods like Gaussian blurring to lower noise and enhance edge detection capabilities.
Laplacian: The second-order derivative nature of the Laplacian operator makes it sensitive to noise, which may increase noise in the image.The sum of second-order partial derivatives in both the x and y directions is calculated using the Laplacian operator, which is a second-order differential operator. Sobel: Sobel operators have a lower noise threshold, particularly when used in combination with smoothing methods like Gaussian blur, which lowers noise in gradient pictures.The gradient of the image intensity function in the x and y directions is independently computed using Sobel operators, which are first-order differential operators.
Hue (H): Hue is a representation of the color. It's frequently defined as the color's dominant light wavelength. Hue is commonly expressed as an angle around a color wheel in the HSV color system, with values ranging from 0 to 360 degrees. The values are typically standardized to fall between 0 and 1. The normal locations of red, green, and blue are 0 degrees, 120 degrees, and 240 degrees, respectively, with various hues scattered in between. Saturation (S): The intensity or brightness of the color is represented by saturation. Higher saturation numbers denote more vibrant, intense colors, whereas a saturation value of 0 represents a shade of gray. Typically, saturation is shown as a percentage, where 100% indicates full saturation and 0% represents total desaturation (gray). Value (V): Value is a representation of the color's lightness or brightness. Higher numbers denote brighter tones of the color; a value of 0 represents black. A percentage is another common way to express value, where 100% is the fully lighted color and 0% is black.
Manually downloaded photos of huamns and not humans from google. Created one folder named image processing in that folder created 3 sub folders namely training , test and validate in these folders there are subfolders containing human and not human image. The not humans consist of table, animal, flowers etc.Convolutional neural networks (CNNs) are trained using the TensorFlow/Keras script to categorize images into two groups: "Human" and "Not Human."